Completing the implementation of colocated sites.

[ted-ross]

No description provided.

[JPadovano1483]

@ted-ross @fgiorgetti During the initial deployment of the colocated site, I get logs showing an attempt to create the vms-colo-manage secret twice, failing on the second attempt since it already exists. Have either of you run into this?

[fgiorgetti]

@ted-ross @fgiorgetti During the initial deployment of the colocated site, I get logs showing an attempt to create the vms-colo-manage secret twice, failing on the second attempt since it already exists. Have either of you run into this?

@JPadovano1483 I am not seeing it.
Did you update your ClusterRole adding permission to manage listeners?
I remember that without it, I was seeing some 403 errors during the attempt to create the Listener resource.
But other than that, I am not seeing errors on mgmg controller and on site-ctrl logs.

[JPadovano1483]

@ted-ross @fgiorgetti During the initial deployment of the colocated site, I get logs showing an attempt to create the vms-colo-manage secret twice, failing on the second attempt since it already exists. Have either of you run into this?

@JPadovano1483 I am not seeing it. Did you update your ClusterRole adding permission to manage listeners? I remember that without it, I was seeing some 403 errors during the attempt to create the Listener resource. But other than that, I am not seeing errors on mgmg controller and on site-ctrl logs.

Yes the clusterrole is updated. Hmm ok, just me then. Haha I'll keep digging then and see if I can find whats going on for me.

[JPadovano1483]

After some more testing, I am getting a mixed bag of results. Sometimes everything deploys without any duplication errors. Other times, it tries to recreate almost all the resources. I think this is because there are several paths that call the "visitNamespace" function, which is causing a race condition.

[ted-ross]

After some more testing, I am getting a mixed bag of results. Sometimes everything deploys without any duplication errors. Other times, it tries to recreate almost all the resources. I think this is because there are several paths that call the "visitNamespace" function, which is causing a race condition.

I think this is right. If there are multiple instances of visitNamespace running concurrently, we would probably see what James is seeing.